Open-source realtime platform and Agents framework for voice AI.
Fish Audio
Text-to-speech and voice cloning with open models.
SharkScore
#4 in Audio & voice
Why it scores 68.1
Fish Audio has a SharkScore of 68.1/100, ranked #4 of 225 in Audio & voice (updated 2026-11-19).
Each part compares Fish Audio with other listings in Audio & voice. Evidence covers 87% of the score's weight.
- Adoption
- 47 · 30%
- User reviews
- no data · 25%
- Maintenance
- 100 · 20%
- Trust & transparency
- 80 · 15%
- Momentum
- no data · 10%
Paying never changes this score. How ranking works · method v2026-11-18.2
RankShark review of Fish Audio
Automated evaluation by RankShark · 18 Nov 2026 · method v1.0 · How we review
We were not paid for this review.
As of 2026-11-18, Fish Audio provides voice generation and related audio tools, with a free tier and a separate pay-as-you-go API. [E1, E2, E3]
Fish Audio offers text-to-speech, speech-to-text, voice cloning, and other audio tools. Its developer API supports speech generation and transcription, with REST and WebSocket options. [E1, E3]
Best for
Not for
- Free-plan users who need commercial-use rights; the pricing page limits the free plan to personal, non-commercial projects. · home page, pricing page
- Teams requiring an API free tier with an SLA; the vendor says S2.1 Pro Free has no SLA. · docs
Key features
- Text-to-speech. The vendor says its text-to-speech supports emotion and direction tags, and lists generation limits of up to 30,000 characters per generation on Pro. · home page, pricing page, docs
- Voice cloning. The vendor says voice cloning is available, including IVC from 30 seconds of audio through its developer offering. · home page, docs
- Speech-to-text. The vendor says its transcription supports multiple speakers and emotion tags; its developer page lists the Transcribe-1 model. · home page, docs
- Voice library. The vendor says its library contains more than 2,000,000 voices. · home page
- Developer API. The vendor says the API supports REST and WebSocket streaming, with Python and TypeScript SDKs. · docs
- Self-hosting. The vendor says its open-weight models can be deployed in a VPC, data center, sovereign cloud, or air-gapped environment as an Enterprise engagement. · docs
Pros
- The free subscription tier includes monthly credits and generation time. · pricing page
- The API documentation lists REST, WebSocket streaming, and official Python and TypeScript SDKs. · docs
- The product covers several audio tasks, including text-to-speech, speech-to-text, voice cloning, and audio translation. · home page
Cons
- The free subscription tier has a 500-character per-generation limit and does not include commercial use. · home page, pricing page
- The vendor says requests to the free S2.1 Pro API may be used to improve model quality, and that the free API has no SLA. · docs
- No release in the last 180 days (latest: v1.5.1 on 2025-05-31).
Pricing
| Plan | Price | Limits |
|---|---|---|
| Free Tier | $0/mo | 8,000 credits monthly; up to 7 minutes generation; up to 500 characters per generation; 3 public voice slots; standard generation speed. Commercial use is not included. |
| Plus | $11/mo billed annually ($132 billed annually); $15/mo shown | 250,000 credits monthly; up to 200 minutes generation; up to 15,000 characters per generation; unlimited public + 10 private voice slots; 1 professional voice slot; commercial use allowed. |
| Pro | $75/mo billed annually ($900 billed annually); $100/mo shown | 2,000,000 credits monthly; up to 1,620 minutes generation; 3 team seats included; up to 30,000 characters per generation; unlimited voice slots; 5 professional voice slots. |
Checked 18 Nov 2026 · source
- Free plan
- 8,000 credits monthly; up to 7 minutes of generation; up to 500 characters per generation; 3 public voice slots. Personal, non-commercial use only. · home page, pricing page
- Free trial
- Not stated (checked 18 Nov 2026)
- Trains on customer data
- For S2.1 Pro Free API requests, the vendor says requests may be used to improve model quality. Other customer-data training practices are not stated in the supplied evidence. · docs
- Data retention
- Not stated (checked 18 Nov 2026)
- Data regions
- Not stated (checked 18 Nov 2026)
- Data processing agreement
- Not stated (checked 18 Nov 2026)
- AI models used
- The developer page lists S2.1 Pro Free, S2.1 Pro, and Transcribe-1. It also names the open-weight model families fish-speech, S1, and S2. · docs
- API
- API access is available with pay-as-you-go pricing. The developer page lists REST and WebSocket streaming. It gives S2.1 Pro pricing as $15 per 1 million UTF-8 bytes and Transcribe-1 as $0.36 per hour. · pricing page, docs
- SDKs
- Official Python and TypeScript SDKs. · docs
- Webhooks
- Not stated (checked 18 Nov 2026)
- Platforms
- Not stated (checked 18 Nov 2026)
- Security claims
- The vendor says its open-weight models can be self-hosted in a VPC, data center, sovereign cloud, or air-gapped environment. · docs
What we measured
- GitHub stars · now
- 33,768
- Latest release · now
- v1.5.1 on 2025-05-31
- Tranco traffic rank · list of the day
- #13,468
How we tested · sources
- Hourly uptime and response-time checks of the product's URL.
- Repository, release and download data from GitHub, npm, PyPI and Hugging Face where they exist.
- We did not sign up, pay, or use the product as a customer.
- home page: https://fish.audio/ · fetched 18 Nov 2026
- pricing page: https://fish.audio/plan/ · fetched 18 Nov 2026
- docs: https://fish.audio/developers/ · fetched 18 Nov 2026
- privacy page: https://fish.audio/privacy/ · fetched 18 Nov 2026
- terms: https://fish.audio/terms/ · fetched 18 Nov 2026
Something wrong? Report an error with a public link that shows it; corrections re-run the review.
Top alternatives to Fish Audio
See all- 72.9ToolOpen source · from $50/mo
- 72.5
Lifelike voices, dubbing and voice agents.
ToolFreemium · from $6/mo - 69.3
Speech-to-text, text-to-speech and voice agent APIs.
ToolFreemium · from $333.33/mo - 67.3
Text-to-speech reader for documents and web pages.
ToolFreemium · from $29/mo